If you are thinking of pursuing a PhD in computer architecture with me, 
the following comments may be useful:

* I believe that one of the most important decisions facing you in your road 
to a PhD is your choice of PhD advisor.  It is very important to work with
someone you are on the same wavelength with, and it must be hell to work
with someone who you dread spending time with.  If you are thinking of
working with me, please check with other students who have interacted
with me, either in classes or on exams or in my research group.  Other
students are a far better source of information on what I am all about,
in my view, than just about anything else.  They will give you a better 
sense of what I consider important, how easy/difficult I am to get along 
with, how hard I make students (and myself) work, etc.

* I currently have 12 graduate students, so I am not eagerly looking for
more.  However, I am always interested in working with students who are 
very bright, highly motivated, and on my wavelength.  The best way to 
demonstrate all of this is to take a course with me so we will both know
whether it makes sense for us to try to work together.  If we were to decide 
to work together without first having had such a substantive interaction, 
we both would be inviting disaster.  So, I will not do that.  It generally 
takes 5 years or more to complete the PhD, and it makes no sense for both 
of us to be miserable for that long a period of time.

* If you are going to do serious work in microarchitecture (my specialty),
it is important to have a solid grounding in the fundamentals.  Actually,
regardless what you want to do serious work in, it is important to have
a solid grounding in the fundamentals.  In computer architecture, this 
starts with a solid hard-core senior level undergraduate course in computer
architecture.  If you have not had the equivalent of EE 360N, as described
in my syllabi accessible via my home page, I strongly suggest you take it
at UT.  From my experience, most students have not had a course equivalent
to 360N.  Feel free to talk to me further about this if you wish. 

* You should also have a solid undergraduate foundation in vlsi circuit 
design (360R at UT), and courses in Operating Systems and Compiler Design.  
Computer Architecture is the interface between the hardware and the software.
Knowing more about what needs to execute optimally and knowing more about
how the structures in a microarchitecture get implemented as circuits will
make your research more relevant.  It will also make you a better 
microarchitect. 

Whether or not you and I end up working together on your PhD, I wish you
good luck in your time here at UT.

Yale Patt